The goal of the proposed research is to investigate predictors, patterns and outcomes of blood pressure (BP) management among patients with diabetes and coexisting health conditions, in order to use this information to optimize existing BP guidelines. We will use a linked registry-claims database to investigate the pattern of BP control by comorbidities, assess the relation between BP and adverse outcomes, and develop guideline prototypes for BP management in diabetes patients with complex health status. The Specific Aims are: (1) To assess the effect of specific vascular and nonvascular comorbidities on BP patterns over time in patients with type 2 diabetes mellitus. We hypothesize that trajectories of BP will be different in diabetes patients who have specific coexisting conditions and different management interventions. (2). To evaluate the association between comorbidity burden, BP control and changes in BP level, and adverse clinical outcomes (stroke, ischemic cardiac events, hospitalization, and death) in diabetes patients. We hypothesize that complex health status, rather than BP level, will more strongly predict adverse events in diabetes patients with multiple comorbidities and vulnerabilities. For those patients without a complex health status, however, we hypothesize that BP level will be a predictor of adverse outcomes. (3) Using the results from Specific Aims 1 and 2, to develop guideline prototypes that minimize the probability of adverse events in complex patients. We will use information regarding the trajectories of BP for specific comorbidities and data from the regression models which predict adverse clinical outcomes, to generate the probabilities of incorporating additional data elements into existing BP guidelines for complex diabetes patients. This is an analytic epidemiology study. It will use secondary data analysis of a clinical diabetes registry containing measures of diabetes management (HbA1c, LDL level, BP level, among others) that will be linked to claims from a Health Maintenance Organization and Medicare. Medicare claims are available through the University of Michigan's participation in the Physician Group Practice (PGP) Medicare Demonstration Project. This project assigns UM a population of fee-for service Medicare patients cared for at UM;most Medicare patients in the diabetes registry are assigned. Participants will be the approximately 20,000 registry patients from 2005-2008, approximately 70% of whom will have claims and multiple BP measures on the registry. Using this linked registry data, this research will investigate blood pressure (BP) level among patients with diabetes from middle to advanced old age, with different comorbidity burdens. Major outcomes include stroke, new ischemic cardiac event, death, and increased hospitalization. Covariates of most interest as predictors of both BP level and clinical outcomes in diabetes are coronary artery disease, complex health status (defined as comorbidity burden) and age. Results of this research will provide practical clinical evidence for modifying BP management in patients with diabetes and complex health status.